Environmental DNA metabarcoding for biodiversity monitoring of a highly diverse tropical fish community in a coral reef lagoon: Estimation of species richness and detection of habitat segregation

An environmental DNA (eDNA) metabarcoding approach has been widely used for biodiversity monitoring of fishes, although it has rarely been applied to tropical and subtropical aquatic ecosystems, where species diversity is remarkably high. This study examined the extent to which species richness can be estimated in a small coral reef lagoon (1,500 × 900 m) near Okinawa Island, southern Japan, where the surrounding waters are likely to harbor more than 1,500 species of fish. During 2015–2017, a total of 16 capture‐based surveys were conducted to create a faunal list of fish species, followed by eDNA metabarcoding based on seawater samples taken from 11 sites in the lagoon on a day in May 2019. We also tested whether eDNA metabarcoding could detect differences between adjacent fish communities inhabiting the offshore reef edge and shore‐side seagrass beds within the lagoon. A total of 217 fish species were confirmed by the capture‐based samplings, while 291 fish species were detected by eDNA metabarcoding, identifying a total of 410 species distributed across 119 families and 193 genera. Of these 410 species, only 96 (24% of the total) were commonly identified by both methods, indicating that capture‐based surveys failed to collect a number of species detected by eDNA metabarcoding. Interestingly, two different approaches to estimate species richness based on eDNA data yielded values close to the 410 species, including one that suggested an additional three or more eDNA surveys from 11 sites (36 samples) would detect 90% of the 410 species. In addition, nonmetric multidimensional scaling for fish assemblages clearly distinguished between the fish communities of the offshore reef edge and those of the shore‐side seagrass beds. This study demonstrates that an eDNA metabarcoding approach is useful for estimating species richness and detection of habitat segregation even in ecosystems with remarkably high species diversity.